Mapping Motor Vehicle Collisions in NYC

This section visualizes the geographic distribution of collisions across New York City.

âš  Data Note

We use latitude and longitude directly from the cleaned dataset.

1. Basic Leaflet Map (5,000 samples)

set.seed(123)

mvc_sample <- mvc_clean |>
  filter(!is.na(latitude), !is.na(longitude)) |>
  slice_sample(n = 5000)

leaflet(mvc_sample) |>
  addProviderTiles(providers$CartoDB.Positron) |>
  setView(lng = -73.97, lat = 40.75, zoom = 11) |>
  addCircleMarkers(
    lng = ~longitude,
    lat = ~latitude,
    radius = 3,
    stroke = FALSE,
    fillOpacity = 0.5,
    color = "#2C93E8"
  )

2. Severity Map (Property Damage / Injury / Fatal)

mvc_sample2 <- mvc_clean |>
  mutate(
    severity = case_when(
      number_of_persons_killed > 0 ~ "Fatal",
      number_of_persons_injured > 0 ~ "Injury",
      TRUE ~ "Property Damage Only"
    )
  ) |>
  slice_sample(n = 5000)

pal <- colorFactor(
  palette = c("#7db7ff", "#2C93E8", "#FF5733"),
  domain = c("Property Damage Only", "Injury", "Fatal")
)

leaflet(mvc_sample2) |>
  addProviderTiles(providers$CartoDB.Positron) |>
  setView(lng = -73.97, lat = 40.75, zoom = 11) |>
  addCircleMarkers(
    lng = ~longitude,
    lat = ~latitude,
    radius = 3,
    stroke = FALSE,
    fillOpacity = 0.6,
    color = ~pal(severity),
    label = ~htmlEscape(severity)
  ) |>
  addLegend(
    position = "bottomright",
    pal = pal,
    values = mvc_sample2$severity,
    title = "Severity"
  )

3. Heat Map (Collision Density)

leaflet(mvc_sample) |>
  addProviderTiles(providers$CartoDB.DarkMatter) |>
  setView(lng = -73.97, lat = 40.75, zoom = 11) |>
  addHeatmap(
    lng = ~longitude,
    lat = ~latitude,
    blur = 20,
    max = 0.05,
    radius = 10
  )

Summary

Spatial patterns reveal:

  • Hotspots in Manhattan, Brooklyn, Western Queens
  • Severity clustering along major roadways
  • Borough-level differences in spatial spread

Mapping provides vital context for subsequent statistical modeling.